Dear Income Futurist,

Say you ask Muse to keep an eye on hotel prices for a trip next spring, then you shut the laptop and go to bed.

Muse keeps working.

Every Muse user has a dedicated virtual computer in Meta’s cloud, with its own processor, memory and browser, and it runs scheduled jobs long after you’ve walked away. So at 3 a.m., while you’re out cold, a machine in a Meta data center is still checking rates for you.

OpenAI made the same move at the end of September. Its always-on agents in ChatGPT each run on a cloud computer of their own, too. Now picture hundreds of millions of people doing what you just did, each one leaving a computer running all night.

That’s a mountain of hardware, and it cuts against a story a lot of investors are telling themselves right now.

Part one of this series showed the same level of AI performance getting about 13x cheaper every year, and plenty of people read that as a sign the spending boom has to cool off.

If the work keeps getting cheaper, the bill should shrink, right?

Britain ran into that exact question with coal. In 1865, an economist named William Stanley Jevons noticed that steam engines kept getting more efficient, and Britain kept burning more coal anyway, because every jump in efficiency made some new use worth the fuel. Cheaper AI works the same way: nobody would leave a computer running all night to watch hotel prices if it still cost what it did two years ago.

So who gets paid when all those overnight computers are humming? I sort it into three groups.

The front door belongs to the company that owns your habit, and in this series that’s mighty Meta Platforms (META).

Follow that hotel search down a level and you hit the pipes. Somebody has to supply the chips, the memory and the power to keep it running around the clock. Amazon.com (AMZN) raised its 2026 investment plan to $220 billion, pointed to pricier memory, and said it still can’t meet AI demand.

Mark Zuckerberg said on Meta’s last earnings call that the company is “getting a lot of offers for compute at a significant premium over what we paid for it.” When the biggest buyers of compute are fielding premium offers for their own capacity, that’s scarcity. The memory makers, the grid builders and the security firms that check every bot’s ID all live down here. When Muse showed up at Amazon without identifying itself, Amazon blocked it, and now every website needs to know which bot is knocking and whether to let it in.

The middlemen are the ones I’d leave.

Think about how that hotel booking used to go: you scroll, you compare, you see “only 2 rooms left” and you panic-book. Your agent won’t care about the banner. Goldman Sachs’ trading desk has a name for these businesses: “consumer inertia,” companies that profit because customers stay put out of habit, from insurers and phone carriers to streaming services and travel sites. On September 22, banks, insurers and travel names slid on Muse fears, with Charles Schwab (SCHW) and Allstate (ALL) each down more than five percent. The next day, Booking Holdings (BKNG) fell five percent, and Expedia Group (EXPE), which had just signed on to work with Muse, dropped nearly eight percent.

Meta’s the hard one for me. It owns the door, but it’s also writing enormous checks: $130 to $145 billion in planned capital spending this year, with analysts expecting free cash flow to turn negative. When a company spends that hard, I’d rather trade it with defined risk (you know your maximum loss going in) than marry the stock.

Long-dated calls, or LEAPS, give you that, and in a choppy, high-volatility market the strike you pick matters. In-the-money calls around 80 delta (they move about 80 cents for every dollar the stock moves) behave mostly like stock, and the most you can lose is what you paid. Out-of-the-money calls work best when a stock breaks out cleanly. When it chops sideways, theta (time decay) and falling volatility eat away at them, and they can lose value even when the story’s right.

Watch the calendar, too. Meta reports third-quarter results in late October, the first real look at Muse in the numbers. Short-dated calls bought right before earnings carry extra volatility in the price, and that volatility usually comes out of the options right after the report, so I’d rather own time than a lottery ticket.

Buy scarcity, avoid abundance. Intelligence just turned into the abundant thing, so I want to own the front door and the pipes it runs through.

I hope you’ve enjoyed this little series.

Here for a good time… AND a long time,

Hans